Cytosolic phospholipase A2 and eicosanoids modulate life, deathand function of human osteoclasts in vitro$

Hugues Allard-Chamard, Philippe Dufort, Sonia Haroun, Artur J. de Brum-Fernandes n

Division of Rheumatology, Faculté de médecine et des sciences de la santé de l0 Université de Sherbrooke et Centre de Recherche Clinique Étienne-Le Bel,Sherbrooke, Québec, Canada

a r t i c l e i n f o

Article history:Received 9 September 2013Received in revised form16 December 2013Accepted 17 December 2013

Keywords:OsteoclastsBoneProstaglandinsEicosanoidsPhospholipases

a b s t r a c t

Introduction: Eicosanoids are important in bone physiology but the specific function of phopholipase enzymeshas not been determined in osteoclasts. The objective of this is study was to determine the presence of cPLA2 inhuman in vitro-differentiated osteoclasts as well as osteoclasts in situ from bone biopsies.Materials and methods: Osteoclastogenesis, apoptosis, bone resorption and the modulation of actin cytoske-leton assays were performed on osteoclasts differentiated in vitro. Immunohistochemistry was done indifferentiated osteoclasts as well as on bone biopsies.Results: Human osteoclasts from normal, fetal, osteoarthritic, osteoporotic and Pagetic bone biopsies expresscPLA2 and stimulation with RANKL increases cPLA2 phosphorylation in vitro. Inhibition of cPLA2 increasedosteoclastogenesis and decreased apoptosis but decreased the capacity of osteoclasts to generate actin ringsand to resorb bone.Discussion and conclusions: These results suggest that cPLA2 modulates osteoclast functions and could be auseful target in bone diseases with hyperactivated osteoclasts.

& 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Eicosanoid biosynthesis is highly regulated and depends upon theactivation of phospholipases A2 (PLA2s) and the subsequent liberationof arachidonic acid (AA) sequestered within the membrane phospho-lipids [1]. Enzymes responsible for this PLA2 activity have been dividedinto four basic families: cytosolic phospholipases A2, calcium-independent phospholipases A2, secreted phospholipases A2 andlipoprotein-associated phospholipases A2. The cytosolic PLA2 (cPLA2)family consists of six enzymes (cPLA2α, β, γ, δ, ε and ζ), among whichcPLA2α is the best studied [2]. Its intracellular regulation and activationis unique among all the PLA2 family members because of its highselectivity for arachidonic acid-containing phospholipids substratesand its tight regulation by Ca2þ , phosphatidylinositol 4,5-bispho-sphate and phosphorylation on serines 505 and 515 [3–6]. The freeAA generated by fully activated phosphorylated-cPLA2α (p-cPLA2) canbe oxygenated into a variety of lipidic mediators, collectively called

eicosanoids. These compounds are involved in a broad spectrum ofphysiologic and pathophysiologic reactions as shown by their involve-ment in arthritis [1], osteoarthritis [7], asthma [8], normal renal [9]and female reproductive function [10], among others. Human osteo-clasts from healthy subjects present cPLA2 activity and produceprostaglandins [11] which are, in turn, known for their great impacton osteoclast functions by suppressing [12] or inducing [13] osteo-clastogenesis, bone resorption and apoptosis [13–15]. In fact, ourlaboratory has already extensively studied the effect of prostaglandinsin OCs metabolism and the receptors involved in their actions[11,12,16]. However, the role of osteoclastic PLA2, the primary step oftheir biogenesis pathway, has never been clarified due to the complex-ity of single cell approaches.

In this study, our goals were to reveal the role of cPLA2 in OCsmetabolism for a better understanding of OC physiology and toelaborate new pharmacological targets to interfere with OC func-tions. We focus on determining, using immunostaining, if humanosteoclasts express cPLA2 in healthy as well as in pathologic boneand to study its possible role in the modulation of osteoclastbiology, essentially osteoclastogenesis, apoptosis, bone resorptionand cytoskeleton assembly.

2. Materials and methods

2.1. Reagents

FBS was purchased from Gibco (Invitrogen Canada), Macrophage-colony stimulating factor (M-CSF) was from MJS BioLynx, Canada.

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/plefa

Prostaglandins, Leukotrienes and EssentialFatty Acids

0952-3278/$ - see front matter & 2014 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.plefa.2013.12.009

☆This study was support by grants from the Canadian Institutes of HealthResearch. H. Allard-Chamard received a scholarship from the Canadian Institutesof Health Research and the Fonds de Recherche en Santé du Québec.

n Correspondence to: Division of Rheumatology, Faculté de médecine et desSciences de la santé de l0 Université de Sherbrooke et Centre de Recherche CliniqueÉtienne-Le Bel, 3001-12e Avenue Nord, Sherbrooke, Québec, Canada J1H 5N4.Tel.: þ819 564 5417; fax: þ819 564 5265.

E-mail addresses: hugues.allard-chamard@usherbrooke.ca (H. Allard-Chamard),philippe.dufort@usherbrooke.ca (P. Dufort),sonia.haroun@usherbrooke.ca (S. Haroun),artur.fernandes@usherbrooke.ca (A.J. de Brum-Fernandes).

Prostaglandins, Leukotrienes and Essential Fatty Acids 90 (2014) 117–123

Rhodamine-conjugated phalloidin, To-Pro-3 iodine and chickenanti-goat secondary antibody coupled to Alexa 488 were purchasedfrom Molecular Probes (Invitrogen Canada). Vectastain ABC, DABsubstrate kits and Vectashield were from Vector Laboratories. ThecPLA2 inhibitor methyl arachidonyl fluorophosphonate (MAFP) [17],Arachidonyl trifluoromethyl ketone (AACOCF3) [18] and compoundCAY10502 [19] were from Cayman Chemical (Ann. Arbor, MI, USA).Pyrrophenone, another cPLA2 inhibitor [20], was kindly provided byDr. Takaski Ono (Shionogi Research Laboratories). All cPLA2 inhibi-tors where dissolved in ethanol. The goat anti-human cPLA2 poly-clonal antibody was from Santa Cruz Biotechnology (Santa Cruz, CA,USA). The rabbit anti-goat biotinylated antibody was from Calbio-chem (Merck, Germany). The anti-CD14 magnetic micro-beads werefrom Miltenyi Biotec (GmbH, Bergisch Gladbach, Germany). HumanRANKL-glutathione S-transferase (GST) fusion protein was purifiedfrom Escherichia coli strain BL21 and extracted by affinity chromato-graphy as previously described [11]; the rhRANKL-GST plasmid waskindly provided by Dr. M. Manolson, University of Toronto; LPScontamination was absent, as confirmed with the use of E-Toxatetest (kit number ET0200, Sigma-Aldrich). The RANKL-GST biologicalactivity was tested on RAW264.7 cells and shown to be comparable,on a molar basis, to commercially available RANKL (catalog numberT3573, Sigma-Aldrich). Bovine cortical bone was purchased from alocal slaughterhouse, cut in thin slices (150 mm) using a low-speeddiamond saw, devitalized, and used as resorption substrate. All otherreagents were purchased from Sigma-Aldrich Canada, Ltd. (Oakville,ON, Canada).

2.2. Cell culture

To generate mature human osteoclasts in vitro, peripheralblood mononuclear cells (PBMCs) were isolated from blood ofself-reported healthy donors by dextran sedimentation and Ficolldensity gradient. Cells were seeded at a density of 1.5�106 cells/cm2, in 24-well plates on 25 mm sterile glass coverslips or on boneslices. Cells were left to adhere for 24 h in α-modified Eagle’smedium (α-MEM) complemented with 10% FBS and 1% of peni-cillin–streptomycin solution at 37 1C and 5% CO2 humidified atmo-sphere. To stimulate differentiation, M-CSF (10 ng/ml) and RANKL(50 ng/ml) were then added to culture media which was changedsubsequently every three days until 21, when experiments wereconducted. All experimental protocols using human cells wereapproved by the Ethics Review Board of the Centre HospitalierUniversitaire de Sherbrooke.

2.3. Immunostaining of cPLA2 and actin structure

After 21 days of culture the cells were fixed in a solution offormaldehyde 3.7%. Endogenous peroxydase activity was inhibitedby incubation with a solution 0.3% H2O2 and then the cells werepermeabilized for 30 min in a solution of PBS 1% Triton-X 0.3%.Blocking was achieved using permeabilizing solution supplementedwith 0.5% BSA and 10% goat serum. Primary anti-cPLA2 or anti-p-cPLA2 antibody was diluted 1:200 in blocking buffer and incu-bated with the cells for 1 hour. The preparations were rinsed andblocked again for 30 min then secondary antibody diluted 1:200was added and incubated for 1 h. Cells were then treated with theVectastain peroxidase standard ABC kit and revealed using theperoxidase substrate kit DAB, both according to the manufacturer’sprotocols. Immunofluorescence was performed using the sameprotocol but cells were fixed using paraformaldehyde 4%, thesecondary antibody was Alexa 488 and the counterstain wasrhodamine-conjugated phalloidin 0.3 mM, an F-actin stain, andbisbenzimide H 33342 0.26 ng/ml and To-Pro-3 1 mM both nuclearmarkers. Immunohistochemistry for cPLA2 was performed onparaffin-embedded fetal, osteoporotic, osteoarthritic, Pagetic and

normal adult bone tissue samples, as described elsewhere [16]. Fetalbone specimens were embedded in paraffin by the local pathologyservice. Normal and pathologic adult bone tissue samples weredonated by Dr. M. Alini of the Orthopedic Research Laboratory,McGill University, Montreal. Specimens were observed using bright-field microscopy or BioRad MRC 1024 confocal microscope forfluorescence.

2.4. Osteoclastogenesis assay

For osteoclastogenesis assays, PBMCs were differentiated for 21days in presence of M-CSF (10 ng/ml) and RANKL (50 ng/ml). cPLA2

inhibitors were added to culture media for various amounts oftime. Differentiated cells were fixed and washed with PBS. Thepresence of tartrate-resistant acid phosphatase (TRAP) was deter-mined using a leukocyte-acid-phosphatase kit according to themanufacturer’s protocol. Cells were stained for 30 min and coun-terstained with hematoxylin. TRAPþ multinucleated cells contain-ing three or more nuclei were counted as osteoclasts. Results arepresented as the mean of the percentage of cells with osteoclasticphenotype per five random fields of view with a 40� objective.

2.5. Apoptosis assay

The role of cPLA2 in the apoptosis of mature OCs was assessedusing a combination of previously described techniques [21–23].Briefly, mature in vitro-differentiated osteoclasts were incubatedwith CAY10502, a specific cPLA2 inhibitor, arachidonic acid (AA), orboth. The mitochondrial potential (ΔΨm) marker JC-1 was thenadded to discriminate between living and early apoptotic cellsusing the mitochondrial apoptotic pathways [22]. JC-1 peptidepolymerize and precipitate in mitochondria under high ΔΨm, if theΔΨm is low as in early apoptotic, cell JC-1 do not polymerize.Polymerized JC-1 exhibits a bright orange fluorescence and itsunpolymerized form a bright green fluorescence. Concomitantly,cells were treated with the cell impermeable nuclei stain To-Pro-3and bisbenzimide H 33342 at concentration of 5 mM, 1 mM and0.26 ng/ml, respectively, for 30 min. To-Pro-3 staining allows thevisualization of late apoptotic cells or necrotic cells [23] andbisbenzimide H 33342 the visualization of nuclei in living cells [21]to discriminate between OCs that are multinucleated (three or morenuclei) and their precursors.

2.6. Bone resorption assay

For the bone resorption assay PBMCs were plated on devita-lized bovine cortical bone slices and osteoclastogenesis inducedwith M-CSF 10 ng/ml and RANKL 50 ng/ml. After 21 days ofculture, vehicle or cPLA2 inhibitors were added to the media foran extra 9 days in a 37 1C, 10% CO2 humidified atmosphere toacidify culture media and stimulate bone resorption. Bone sliceswere stained with 0.2% toluidine blue for 2 min to reveal lacunarresorption pits. Bone slices were then washed in water, and air-dried before being photographed using brightfield microscopy.Resorption area was quantified using the image analysis programSimplePCI and the results expressed in mm2.

3. Results

3.1. Human osteoclasts express cPLA2

Immunocytochemistry staining (brown using DAB substrate)showed the presence of cPLA2 in osteoclasts differentiated in vitrofrom PBMCs (Fig. 1B), as well as in human osteoclasts in situ innormal adult bone (Fig. 1D), osteoporotic (Fig. 1E), osteoarthritic

H. Allard-Chamard et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 90 (2014) 117–123118

(Fig. 1F), fetal (Fig. 1G) and Pagetic bone biopsies (Fig. 1H). cPLA2

staining was extended to cells surrounding the OCs, indicating thatthis protein is also present in the bone stromal cells. Fig. 1A and Crepresent negative controls without primary antibody. Typicalosteoclasts are shown using full arrow and stromal cell usingarrow head.

3.2. Localization and modulation of phosphorylated-cPLA2 by RANKLin human osteoclasts

Confocal microscopy revealed immunofluorescence staining ofphosphorylated cPLA2 (p-cPLA2) that is mainly localized in thenuclei as shown by the green staining (Fig. 2A) and the corre-sponding fluorogram showing colocalization between p-cPLA2 andthe nuclear marker To-Pro-3 (Fig. 2B). In fact, using confocal micro-scopy with double staining allowed us to show cell compartment

specificity of p-cPLA2 localization. Immunostaining also showedthat the amount of p-cPLA2 could greatly be reduced in serum andcytokine (M-CSF & RANKL) starved cells (Fig. 2D) compared to no-starved controls (Fig. 2C). The decrease in cPLA2 phosphorylationinduced by serum and cytokines starvation could be partiallyreversed by stimulating OCs with RANKL (50 ng/ml) for 15 min(Fig. 2E).

3.3. Effects of cPLA2 inhibition on osteoclastogenesis in vitro.

To assess the effect of cPLA2 inhibition on osteoclastogenesis,PBMCs were induced to differentiate into OCs in the presence ofRANKL 50 ng/ml, M-CSF 10ng/ml and various cPLA2 inhibitors for21 days. All cPLA2 inhibitors tested increased osteoclastogenesis inthis experimental model. As seen in Fig. 3, Pyrrophenone 5 mM,AACOCF3 4.7 mM, CAY10502 2 mM and MAPF 4mM increased

Fig. 1. Expression of cPLA2 in human osteoclasts. Immunocytochemistry was performed on in vitro–differentiated OCs (B), on normal adult bone (D) and slides fromosteoporotic (E), osteoarthritic (F), fetal (G) and Pagetic bone (H). (A) and (C) represent controls without primary antibodies. Full arrows indicate mature OCs and arrowheads indicate stromal cells. Each picture is representative of at least three experiments on different donors samples. Scale bar¼200 mm, magnification 200� .

H. Allard-Chamard et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 90 (2014) 117–123 119

osteoclastogenesis from 82.6723.2 to 201.0729.4 (po0.05),252.2779.0 (po0.001), 233.0768.2 (po0.01) and 269.3729.7(po0.001) multinucleated TRAPþ cells/well respectively.

3.4. Role of cPLA2 inhibition on OCs apoptosis in vitro

Treatment of mature in vitro-differentiated OCs with the cPLA2

inhibitor compound CAY10502 reduced OC apoptosis induced by

serum and cytokines starvation, as shown in Fig. 4. This effect wasdose-dependent, with an IC50 of 1.59�10�7 M. Co-treatment witharachidonic acid 0.25 mM completely abrogated the effect of com-pound CAY10502.

3.5. Effect of cPLA2 in the modulation of bone resorptionand associated actin structure in human OCs.

OCs were differentiated from PBMCs on bovine cortical boneslices to study the effect of cPLA2 in bone resorption. Treatment of

Fig. 2. Localization and modulation by RANKL of phosphorylated-cPLA2 in in vitro-differentiated human osteoclasts. Immunofluorescence, using green Alexa 488 stainingwas performed on in vitro–differentiated OCs and revealed that phosphorylated-cPLA2 is mostly present in OCs nuclei as shown in blue with the nuclei marker To-Pro-3;actin is in red (A). Furthermore, confocal analysis and subsequent fluorogram calculating the intensity of both markers in every pixel confirmed that phosphorylated-cPLA2

and nuclei marker To-Pro-3 colocalize (B). Serum and cytokines deprivation of OCs leads to a decrease in cPLA2 phosphorylation (D) compare to non-starved control (C).RANKL stimulation could partially restore cPLA2 phosphorylation after starvation (E). Each picture is representative of at least three experiments on different donors.Magnification, 600� . (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3. Effects of cPLA2 inhibition on osteoclastogenesis in vitro. PBMCs wereinduced to differentiate into OCs in the presence of RANKL 50 ng/ml and M-CSF10 ng/ml for 21 days. To assess the effect of cPLA2 inhibition on osteoclastogenesis,culture media was supplemented with vehicle or various cPLA2 inhibitors. After 21days of culture, cells were stained for TRAP and counterstained with hematoxylin.TRAPþ cells containing 3 or more nuclei were counted as OCs. Pyrrophenone 5 mM,AACOCF3 4.7 mM, and CAY10502 2 mM and 4 mM MAPF significantly increasedosteoclastogenesis compared to vehicle-treated control cells npo0.05; nnpo0.01;nnnpo0.001 vs control (n¼5).

Fig. 4. Effects of cPLA2 inhibition on OCs apoptosis in vitro. PBMCs were induced todifferentiate into OCs in the presence of RANKL 50 ng/ml and M-CSF 10 ng/ml for 21days then serum and cytokine starved for 24 h in presence of cPLA2 inhibitor, CAY10502,with or without 0.25 mM AA. Cells were then stained with JC-1 mitochondrial potentialmarker, To-Pro-3 and bisbenzimide H 33342 to assess apoptosis and nuclei’s number.Multinucleated cells were counted as OCs, cells that did not precipitate JC-1 and stayedgreen (unprecipitated form of JC-1) or cells that incorporated To-Pro-3 (late apoptoticmarker) and stained red were counted as apoptotic. The results are given in percentageof osteoclasts stained with the apoptosis marker (n¼4).

H. Allard-Chamard et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 90 (2014) 117–123120

fully differentiated OCs during 10 days (days 21–31) with cPLA2

inhibitors reduced resorption area. Pyrrophenone 5 mM, AACOCF34.7 mM, CAY10502 2 mM and MAPF 4 mM decreased resorption areafrom 63681716706 in controls to 1068778235 (po0.0026)887974057 (po0.0015) 1159978026 (po0.0028) 920975017(po0.0017) respectively, as shown in Fig. 5. Actin structures ofglass plated OCs were revealed using rhodamin-phalloidin. Inhibi-tion of cPLA2 with CAY10502 showed a dose-dependent inhibitionof actin ring formation with a calculated IC50 of 1.59�10�7 M. Thisinhibition of actin ring formation could be reversed by cotreat-ment of OCs with AA 0.25 mM. (Fig. 6). Addition of exogenousarachidonic acid competed reverse the effect of CAY10502.

4. Discussion and conclusions

Eicosanoids – more precisely prostaglandins (PGs) – play acrucial role in bone homeostasis as well as in pathophysiology ofbone disease [24,25]. They act on the metabolism of OCs and

modulate their capacity to differentiate and resorb bone, and canact either as stimulators or inhibitors of OC functions, dependingof the PG investigated. Indeed, exogenous PGE2 stimulates osteo-clastogenesis and bone resorption in vitro as well as in vivo [15,26]and exogenous PGD2 is a strong inhibitor of OCs function and apotent inducer of apoptosis [12]. However, the endogenous controlof eicosanoids production by OCs and its contribution to bonemetabolism has received little attention. Our team has previouslyreported that OCs can produce PGs [11], but the implication of thedifferent pathways involved in their biosynthesis as well as thepossible effect of their inhibition had never been demonstrated.

The present study showed that in vitro-differentiated OCs aswell as in in situ OCs from bone biopsies of normal, fetal,osteoporotic, Pagetic and osteoarthritic bone expressed cPLA2,and should thus be able generate free arachidonic acid andprobably to initiate eicosanoid production on their own. Theseresults support our previous findings suggesting that cPLA2 is themost important PLA2 for OCs PLA2 activity [11]. Moreover, theactive form of cPLA2, the p-cPLA2 is present in OCs.

The present research also explored the role of endogenouscPLA2 action in OCs metabolism. Pharmacologic inhibition of cPLA2

with chemically different compounds increased osteoclastogenesisand decreased OC apoptosis by the intrinsic pathway. These resultssuggest that cPLA2 activity is associated with an inhibition ofosteoclastogenesis and with higher levels of OC apoptosis, as hasbeen shown in other experimental models [35,36]. Both effectsshould lead to lower OC activity and reduced bone resorption;they may in fact be linked, as a higher apoptosis rate would lead toa lower number of mature OCs. These results suggest that cPLA2

activity and its subsequent metabolites would exert a negativecontrol of osteoclast function. This is coherent with others findingshowing that prostaglandins inhibit osteoclastogenesis [12,37].Moreover, cyclooxygenase-2 inhibition has been shown to increaseosteoclastogenesis while cyclooxygenase-1 inhibition had noeffect [12] thus highlighting that the net effect of eicosanoidproduction is a decrease of osteoclastogenesis.

Interestingly, inhibition of cPLA2 in mature OCs decreased boneresorption and the number of cells presenting an actin ring,suggesting that cPLA2 activity may be needed for the formationof actin rings and bone resorption. Several eicosanoids have beenshown to have important effects in the modulation of thecytoskeleton of OCs and the formation of actin rings [11,38,39].The sum of current knowledge seems to point toward prostaglan-din acting as chemoattractants [12,40]. Chemokinesis induce byprostaglandins result in modification of OCs actin cytoskeleton andgeneration of lamellipodia and thus inhibits formation of actinring required for resorption. In fact, PGD2 increase lamellipodiathrough their effect on CRTH2 receptor [12] and PGE2 decreaseactin ring trough their EP4 receptor while increasing lamellipodiathrough its EP3 receptor [40]. Our current results shown that atotal depletion of eicosanoids from the cPLA2 pathway lead to OCsthat are no further able to generate mature OCs structure thushighlighting the role of eicosanoids in the control of OCs cytoske-leton. Even though the net result of cPLA2 activity on boneresorption in vivo cannot be extrapolated from the present results,it is interesting to speculate that bone resorption being a down-stream event from osteoclastogenesis, cPLA2 activity could still beassociated with decreased OC activity in vivo.

The effects of cPLA2 activity on OC apoptosis and on thecytoskeleton are probably dependent on the synthesis of eicosa-noids, as the effects of cPLA2 inhibitors could be reversed by theaddition of exogenous arachidonic acid. Furthermore this is alsosuggested by the facts that cyclooxygenase-2 inhibition alsoincreases osteoclastogenesis [11] and that its metabolites inhibitosteoclastogenesis. Other teams using different approaches havealso obtained similar results showing that annexin II stimulates

Fig. 5. Effect of cPLA2 inhibition on bone resorption by in vitro–differentiatedosteoclasts. Differentiated OCs were cultured on bovine cortical bone slices for 21days and stimulated for 10 extra days with cPLA2 inhibitor (Pyrrophenone 5 mM,AACOCF3 4.7 mM, and CAY10502 2 mM and 4 mM MAPF). nnpo0.001 vs control(n¼4).

Fig. 6. Effects of cPLA2 inhibition on mature OCs actin structure in vitro. Mature OCswere stimulated with cPLA2 inhibitor CAY10502 for 2 h and stained with arodamine–phalloidin conjugate to reveal actin structure. Multinucleated cells werecounted as OCs (N¼4).

H. Allard-Chamard et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 90 (2014) 117–123 121

osteoclastogenesis and inhibits prostaglandin production bycPLA2 [34,41].

cPLA2 activity is known to be strictly regulated through post-transcriptional modification [6]. One of these crucial modificationsis cPLA2 phosphorylation on serine 505 (p-cPLA2), as it leads to anincrease in cPLA2 catalytic activity. Our results showed thatexposition of OCs to soluble RANKL has the potential to increaseserine 505 phosphorylation. Indeed, the reduction of phosphor-ylation of serine 505 induced by serum and cytokine starvation ofcultured OCs could be reversed to nearly its baseline level byRANKL treatment. These results suggest that RANKL inducesphosphorylation of cPLA2. In fact, it has been shown that RANKLactivates OCs p38 MAPK which is recognized as a potent phos-phorylator of cPLA2 and could therefore link RANKL to cPLA2

phosphorylation [27]. This is an interesting finding since it shows apotentially new negative feedback of osteoclastogenesis in bonehealth by OCs’ prostaglandins. In fact, RANKL, through the p38MAPK, is able to activate the cPLA2 and induce eicosanoidsproduction by OCs themselves. This means that when activatedby RANKL, OCs may simultaneously activate pathways that sup-press (prostaglandins) and activate (NF-∣B) OCs [12,28]. It ispossible to speculate that this natural anti-OCs’ pathway, throughcPLA2 activation, could prevent extensive bone destruction andstimulate bone formation [28].

It is possible, however, that the role of cPLA2 may not be strictlyrestricted to eicosanoid production. In fact, in our model asubstantial proportion of p-cPLA2 was identified in nuclei, whileunphosphorylated cPLA2 was mostly present in the cytoplasm. Therole of p-cPLA2 in nuclei is still speculative and to our knowledgehas never been described before. It is not yet clear if nucleusp-cPLA2 participates in eicosanoid production since p-cPLA2 is notmembrane bound, a requirement for arachidonic acid release frommembrane-bound phospholipids. Indeed, using confocal micro-scopy, our study showed the colocalization of p-cPLA2 withTo-Pro-3, a DNA marker in cell nuclei [29]. This surprising resultleads us to speculate that p-cPLA2 could act as a scaffoldingprotein. Indeed cPLA2 has already been shown to interact withmany proteins such as annexin-1 [30], annexin-A6 [31], NADPHoxidase [32], integrin αIIbβ3 [33] and annexin II and its subunitp11 [34], and it could therefore act as a structuring protein in thenucleus instead of just participating in eicosanoids’ production.

Briefly, our results show that OCs present cPLA2 and that itsactivity is implicated in the down regulation of osteoclastogenesiswhile increasing OCs apoptosis in vitro. Its activity is also crucial inthe generation of mature and functional cytoskeleton required forbone resorption as inhibition of cPLA2 leads to a reduction informed actin ring and consequently diminished bone resorptionand pit formation. Eicosanoid production is probably implicated inthese effects since arachidonic acid may be used as a rescue drugin apoptosis experiments, but it may not be the only mediator ofcPLA2 effects on OCs. Indeed, p-cPLA2 is surprisingly mostlylocated in the nucleus and exposition of OCs and their precursorsto cPLA2 inhibitors for several days as required for osteoclastogen-esis assay might alter the cell membrane structure and thus exerteffects independent of the eicosanoids pathways. Further studieson the isolated effects of eicosanoids on OC function are needed tounderstand their role in OC biology.

Acknowledgments

This work was supported by grants from the Canadian Insti-tutes of Health Research and the Canadian Arthritis Network. HA-Cwas supported by the FRSQ and CIHR. Pyrrophenone was kindlyprovided by Dr. Takaski Ono (Shionogi Research Laboratories).

References

[1] J. Bonventre, Cytosolic phospholipase A2alpha reigns supreme in arthritis andbone resorption, Trends Immunol. 25 (2004) 116–119.

[2] T. Ohto, N. Uozumi, T. Hirabayashi, T. Shimizu, Identification of novel cytosolicphospholipase A(2)s, murine cPLA(2){delta}, {epsilon}, and {zeta}, which forma gene cluster with cPLA(2){beta}, J. Biol. Chem. 280 (2005) 24576–24583.

[3] J. Casas, M.A. Gijon, A.G. Vigo, M.S. Crespo, J. Balsinde, M.A. Balboa, Phospha-tidylinositol 4,5-bisphosphate anchors cytosolic group IVA phospholipase A2to perinuclear membranes and decreases its calcium requirement for translo-cation in live cells, Mol. Biol. Cell 17 (2006) 155–162.

[4] E.A. Dennis, Phospholipase A2 in eicosanoid generation, Am. J. Respir. Crit.Care Med. 161 (2000) S32–S35.

[5] M.A. Gijon, D.M. Spencer, A.L. Kaiser, C.C. Leslie, Role of phosphorylation sitesand the C2 domain in regulation of cytosolic phospholipase A2, J. Cell Biol. 145(1999) 1219–1232.

[6] Z. Pavicevic, C.C. Leslie, K.U. Malik, cPLA2 phosphorylation at serine-515 andserine-505 is required for arachidonic acid release in vascular smooth musclecells, J. Lipid Res. 49 (2008) 724–737.

[7] E.S. Molloy, G.M. McCarthy, Eicosanoids, osteoarthritis, and crystal depositiondiseases, Curr. Opin. Rheumatol. 17 (2005) 346–350.

[8] J.A. Boyce, Eicosanoids in asthma, allergic inflammation, and host defense,Curr. Mol. Med. 8 (2008) 335–349.

[9] J. Chen, M. Zhao, W. He, et al., Increased dietary NaCl induces renal medullaryPGE2 production and natriuresis via the EP2 receptor, Am. J. Physiol. RenalPhysiol. 295 (2008) F818–F825.

[10] J.D. Godkin, M.P. Roberts, M. Elgayyar, W. Guan, P.K. Tithof, Phospholipase A2regulation of bovine endometrial (BEND) cell prostaglandin production,Reprod. Biol. Endocrinol. 6 (2008) 44.

[11] J.A. Hackett, H. Allard-Chamard, P. Sarrazin, et al., Prostaglandin production byhuman osteoclasts in culture, J. Rheumatol. 33 (2006) 1320–1328.

[12] M. Durand, M.A. Gallant, A.J. de Brum-Fernandes, Prostaglandin D2 receptorscontrol osteoclastogenesis and the activity of human osteoclasts, J. BoneMiner. Res. 23 (2008) 1097–1105.

[13] Z. Li, C. Schem, Y.H. Shi, D. Medina, M. Zhang, Increased COX2 expressionenhances tumor-induced osteoclastic lesions in breast cancer bone metastasis,Clin. Exp. Metastasis 25 (2008) 389–400.

[14] X.Y. Tian, Q. Zhang, R. Zhao, et al., Continuous PGE2 leads to net bone losswhile intermittent PGE2 leads to net bone gain in lumbar vertebral bodies ofadult female rats, Bone 42 (2008) 914–920.

[15] X.Y. Tian, Q. Zhang, R. Zhao, et al., Continuous infusion of PGE2 is catabolicwith a negative bone balance on both cancellous and cortical bone in rats,J. Musculoskelet. Neuronal. Interact. 7 (2007) 372–381.

[16] I. Fortier, M.A. Gallant, J.A. Hackett, C. Patry, A.J. de Brum-Fernandes,Immunolocalization of the prostaglandin E2 receptor subtypes in human bonetissue: differences in foetal, adult normal, osteoporotic and pagetic bone,Prostaglandins Leukot. Essent. Fatty Acids 70 (2004) 431–439.

[17] J. Balsinde, E.A. Dennis, Distinct roles in signal transduction for each of thephospholipase A2 enzymes present in P388D1 macrophages, J. Biol. Chem. 271(1996) 6758–6765.

[18] I.P. Street, H.K. Lin, F. Laliberte, et al., Slow- and tight-binding inhibitors of the85-kDa human phospholipase A2, Biochemistry 32 (1993) 5935–5940.

[19] J. Ludwig, S. Bovens, C. Brauch, A.S. Elfringhoff, M. Lehr, Design and synthesisof 1-indol-1-yl-propan-2-ones as inhibitors of human cytosolic phospholipaseA2alpha, J. Med. Chem. 49 (2006) 2611–2620.

[20] T. Ono, K. Yamada, Y. Chikazawa, et al., Characterization of a novel inhibitor ofcytosolic phospholipase A2alpha, pyrrophenone, Biochem. J. 363 (2002) 727–735.

[21] D.J. Arndt-Jovin, T.M. Jovin, Analysis and sorting of living cells according todeoxyribonucleic acid content, J. Histochem. Cytochem. 25 (1977) 585–589.

[22] A. Cossarizza, M. Baccarani-Contri, G. Kalashnikova, C. Franceschi, A newmethod for the cytofluorimetric analysis of mitochondrial membrane poten-tial using the J-aggregate forming lipophilic cation 5,50 ,6,60-tetrachloro-1,10,3,30-tetraethylbenzimidazolcarbocyanine iodide (JC-1), Biochem. Biophys.Res. Commun. 197 (1993) 40–45.

[23] C.L. Harding, D.R. Lloyd, C.M. McFarlane, M. Al-Rubeai, Using the microcyteflow cytometer to monitor cell number, viability, and apoptosis in mammaliancell culture, Biotechnol. Prog. 16 (2000) 800–802.

[24] Q. Gao, P. Zhan, C.B. Alander, et al., Effects of global or targeted deletion of theEP4 receptor on the response of osteoblasts to prostaglandin in vitro and onbone histomorphometry in aged mice, Bone 45 (2009) 98–103.

[25] W.K. Jung, I.S. Park, S.J. Park, et al., The 15-deoxy-Delta12,14-prostaglandin J2inhibits LPS-stimulated AKT and NF-kappaB activation and suppressesinterleukin-6 in osteoblast-like cells MC3T3E-1, Life Sci. 85 (2009) 46–53.

[26] H. Kaji, T. Sugimoto, M. Kanatani, M. Fukase, M. Kumegawa, K. Chihara,Prostaglandin E2 stimulates osteoclast-like cell formation and bone-resorbing activity via osteoblasts: role of cAMP-dependent protein kinase,J. Bone Miner. Res. 11 (1996) 62–71.

[27] X. Li, N. Udagawa, K. Itoh, et al., p38 MAPK-mediated signals are required forinducing osteoclast differentiation but not for osteoclast function, Endocrinol-ogy 143 (2002) 3105–3113.

[28] Y. Koshihara, M. Kawamura, Prostaglandin D2 stimulates calcification of humanosteoblastic cells, Biochem. Biophys. Res. Commun. 159 (1989) 1206–1212.

[29] L.S. Ploeger, H.F. Dullens, A. Huisman, P.J. van Diest, Fluorescent stains forquantification of DNA by confocal laser scanning microscopy in 3-D, Biotech-nol. Histochem. 83 (2008) 63–69.

H. Allard-Chamard et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 90 (2014) 117–123122

[30] Y.W. Chung, H.Y. Oh, J.Y. Kim, J.H. Kim, I.Y. Kim, Allergen-induced proteolyticcleavage of annexin-1 and activation of cytosolic phospholipase A2 in thelungs of a mouse model of asthma, Proteomics 4 (2004) 3328–3334.

[31] L. Cubells, S. Vila de Muga, F. Tebar, et al., Annexin A6-induced inhibition ofcytoplasmic phospholipase A2 is linked to caveolin-1 export from the Golgi,J. Biol. Chem. 283 (2008) 10174–10183.

[32] Z. Shmelzer, M. Karter, M. Eisenstein, et al., Cytosolic phospholipase A2alpha istargeted to the p47phox-PX domain of the assembled NADPH oxidase via anovel binding site in its C2 domain, J. Biol. Chem. 283 (2008) 31898–31908.

[33] N. Prevost, J.V. Mitsios, H. Kato, et al., Group IVA cytosolic phospholipase A2(cPLA2alpha) and integrin alphaIIbbeta3 reinforce each other0s functionsduring alphaIIbbeta3 signaling in platelets, Blood 113 (2009) 447–457.

[34] A. Bailleux, D. Wendum, F. Audubert, et al., Cytosolic phospholipase A2-p11interaction controls arachidonic acid release as a function of epithelial cellconfluence, Biochem. J. 378 (2004) 307–315.

[35] C.A. Culver, S.M. Michalowski, R.C. Maia, S.M. Laster, The anti-apoptotic effectsof nordihydroguaiaretic acid: inhibition of cPLA(2) activation during TNF-induced apoptosis arises from inhibition of calcium signaling, Life Sci. 77(2005) 2457–2470.

[36] L. Duan, H. Gan, J. Arm, H.G. Remold, Cytosolic phospholipase A2 participateswith TNF-alpha in the induction of apoptosis of human macrophages infectedwith Mycobacterium tuberculosis H37Ra, J. Immunol. 166 (2001) 7469–7476.

[37] K. Ono, H. Kaneko, S. Choudhary, et al., Biphasic effect of prostaglandin E2 onosteoclast formation in spleen cell cultures: role of the EP2 receptor, J. BoneMiner. Res. 20 (2005) 23–29.

[38] O. Dransfeld, I. Rakatzi, S. Sasson, et al., Eicosanoids participate in theregulation of cardiac glucose transport by contribution to a rearrangementof actin cytoskeletal elements, Biochem. J. 359 (2001) 47–54.

[39] S.L. Welles, D. Shepro, H.B. Hechtman, Eicosanoid modulation of stress fibersin cultured bovine aortic endothelial cells, Inflammation 9 (1985) 439–450.

[40] P. Sarrazin, J.A. Hackett, I. Fortier, M.A. Gallant, A. de Brum-Fernandes, Role ofEP3 and EP4 prostaglandin receptors in reorganization of the cytoskeleton inmature human osteoclasts, J. Rheumatol. 31 (2004) 1598–1606.

[41] S. Takahashi, S.V. Reddy, J.M. Chirgwin, et al., Cloning and identification ofannexin II as an autocrine/paracrine factor that increases osteoclast formationand bone resorption, J. Biol. Chem. 269 (1994) 28696–28701.

H. Allard-Chamard et al. / Prostaglandins, Leukotrienes and Essential Fatty Acids 90 (2014) 117–123 123